Agents are ushering in the Antisocial Coding era

AI coding agents are driving a shift toward what the author calls 'Antisocial Coding' — a trend away from collaborative, social development practices. Three key observations: agents create hub-and-spoke communication bottlenecks that make multi-developer collaboration unwieldy; agent-built codebases calcify quickly into overly complex, single-owner projects; and open source projects are closing off contributions (e.g., tldraw removing its test suite to prevent agent-assisted rewrites). The downstream effects include increased bus-factor risk, erosion of collective code ownership, and fewer mentorship opportunities for junior developers. Engineering leaders are urged to consider these organizational impacts before optimizing purely for agent productivity.